**5. Results and discussion**

### **5.1 Influence of substrate concentration and OLR on the COD removal efficiency and operational parameters**

The anaerobic degradability studies were carried out using two different two-phase OMSWs with COD concentrations of 35 g COD/L (OMSW 1) and 150 g COD/L (OMSW 2). The

Influence of Substrate Concentration on the Anaerobic

production rates)

with a COD of 150 g/L.

Degradability of Two-Phase Olive Mill Solid Waste: A Kinetic Evaluation 85

OLR (g COD/(L d)) 3.00 6.01 9.05 12.02 15.03 HRT (d) 50.0 25.0 16.6 12.5 10.0 pH 7.2 7.0 7.0 6.9 6.5

\* *rCH4* (L CH4/(L d)) 0.59 1.13 1.64 2.12 2.05 COD (g/L) 4.80 9.05 12.95 17.50 25.70 Soluble COD 3.05 6.00 8.25 11.30 15.05 VS (g/L) 3.60 6.80 9.70 13.10 19.30

TVFA (g acetic acid/L) 0.56 0.81 1.08 1.25 1.57 Alkalinity (g CaCO3/L) 1.98 1.90 1.81 1.70 1.32 TVFA/Alkalinity 0.23 0.35 0.40 0.61 0.95 Values are the averages of 5 determinations taken over 5 days after the steady-state conditions had been reached. The differences between the observed values were less than 3 % in all cases. (\* *rCH4*: methane

Table 3. Steady-state results under different experimental conditions for the OMSW 2

corresponding to an HRT of 10 days, COD removal was 82.9%.

concentrated substrate used (OMSW 2).

As can be seen in Figure 1 the percentage of COD removed decreased with increased OLR for the two influent substrate concentrations studied. The percentage of COD removal decreased from 93.3% to 83.2% when OLR increased from 0.86 to 4.14 g COD/(L d) for the most diluted substrate (OMSW 1). For the most concentrated influent (OMSW 2) OLRs were varied from 3.00 to 15.03 g COD/(L d) and COD removal efficiencies higher than 88% were obtained at an OLR of 12.02 g COD/(L d). Even under a higher OLR of 15.03 g COD/(L d),

The total effluent CODs of the anaerobic reactor increased with increased OLR for the two influent substrate concentrations studied, as summarized in Tables 2 and 3. Such an increase in the effluent COD was paralleled by a similar increase in the effluent total volatile fatty acids (TVFA). This seems to indicate that, at higher OLR, the effluent total COD and mainly soluble COD is largely composed of the unused volatile acids produced in the reactor.

Given that the buffering capacity of the experimental system was found to be at favourable levels with excessive total alkalinity present at virtually all loadings, the efficiency of the process and the rate of methanogenesis was not very affected. The experimental data obtained in this work indicate that a total alkalinity of about 1.7 g/L as CaCO3 is sufficient to prevent the pH from dropping to below 7.0 at an OLR of 9.05 g COD/(L d) for the most

The pH in the reactor was always higher that 7.0 for all the HRTs and OLRs studied corresponding to the most diluted OMSW studied. In addition, pH values equal or higher than 6.9 were observed for OLRs lower than 12.02 g COD/(L d) and HRTs higher than 12.5 d when the most concentrated influent was processed, with pH of 7.2 as a maximum value achieved. This high stability can be attributed to carbonate/bicarbonate buffering. This is produced by the generation of CO2 in the digestion process which is not completely removed from the reactor as gas. Buffering in anaerobic digestion is normally due to bicarbonate, as carbonate is, generally, negligible if compared to the bicarbonate (carbonate/bicarbonate ratio is equal to 0.01 for pH 8.2) (Speece, 1983). The buffering guards

experiments were performed using progressive influent substrate concentrations, those corresponding to the OMSW 1 being the first ones and those corresponding to the OMSW 2 carried out at the end of the study.

Tables 2 and 3 summarize the steady-state operating results including HRT, OLR, methane production rates (*rCH4*), total and soluble CODs, VS, TVFA, alkalinity and TVFA/alkalinity ratio for the OMSW 1 and OMSW 2, respectively (Borja et al., 2002).

Figure 1 shows the variation of the COD removal efficiency with the OLR for the two OMSWs used.

Fig. 1. Variation of the percentage of COD removed with the OLR for the two OMSWs used (■: OMSW 1; ●: OMSW 2).


Values are the averages of 5 determinations taken over 5 days after the steady-state conditions had been reached. The differences between the observed values were less than 3 % in all cases. (\**rCH4*: methane production rates)

Table 2. Steady-state results under different experimental conditions for the OMSW 1 with a COD of 35 g/L.

experiments were performed using progressive influent substrate concentrations, those corresponding to the OMSW 1 being the first ones and those corresponding to the OMSW 2

Tables 2 and 3 summarize the steady-state operating results including HRT, OLR, methane production rates (*rCH4*), total and soluble CODs, VS, TVFA, alkalinity and TVFA/alkalinity

Figure 1 shows the variation of the COD removal efficiency with the OLR for the two

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 **OLR (g COD/(L d))**

Fig. 1. Variation of the percentage of COD removed with the OLR for the two OMSWs used

OLR (g COD/(L d)) 0.86 1.21 1.38 1.72 2.08 2.76 3.45 4.14

\* *rCH4* (L CH4/(L d)) 0.24 0.34 0.38 0.47 0.56 0.73 0.91 0.85 COD (g/L) 2.30 2.50 2.74 3.40 3.85 4.20 4.50 5.80 Soluble COD 0.72 1.20 1.40 1.65 1.90 2.15 2.35 3.80 VS (g/L) 1.70 1.88 2.07 2.40 2.75 3.10 3.40 4.50

TVFA (g acetic acid/L) 0.105 0.155 0.180 0.205 0.215 0.260 0.310 0.495

CaCO3/L) 1.950 1.850 1.715 1.690 1.640 1.690 1.670 1.410

TVFA/Alkalinity 0.04 0.07 0.09 0.10 0.11 0.13 0.15 0.29

Values are the averages of 5 determinations taken over 5 days after the steady-state conditions had been reached. The differences between the observed values were less than

Table 2. Steady-state results under different experimental conditions for the OMSW 1

HRT (d) 40.0 28.6 25.0 20.0 16.6 12.5 10.0 8.3 pH 7.9 7.8 8.0 7.9 8.0 7.9 7.8 7.1

ratio for the OMSW 1 and OMSW 2, respectively (Borja et al., 2002).

carried out at the end of the study.

OMSWs used.

82.0 84.0 86.0 88.0 90.0 92.0 94.0 96.0 98.0

(■: OMSW 1; ●: OMSW 2).

Alkalinity (g

with a COD of 35 g/L.

3 % in all cases. (\**rCH4*: methane production rates)

**COD removed (%)**


Values are the averages of 5 determinations taken over 5 days after the steady-state conditions had been reached. The differences between the observed values were less than 3 % in all cases. (\* *rCH4*: methane production rates)

Table 3. Steady-state results under different experimental conditions for the OMSW 2 with a COD of 150 g/L.

As can be seen in Figure 1 the percentage of COD removed decreased with increased OLR for the two influent substrate concentrations studied. The percentage of COD removal decreased from 93.3% to 83.2% when OLR increased from 0.86 to 4.14 g COD/(L d) for the most diluted substrate (OMSW 1). For the most concentrated influent (OMSW 2) OLRs were varied from 3.00 to 15.03 g COD/(L d) and COD removal efficiencies higher than 88% were obtained at an OLR of 12.02 g COD/(L d). Even under a higher OLR of 15.03 g COD/(L d), corresponding to an HRT of 10 days, COD removal was 82.9%.

The total effluent CODs of the anaerobic reactor increased with increased OLR for the two influent substrate concentrations studied, as summarized in Tables 2 and 3. Such an increase in the effluent COD was paralleled by a similar increase in the effluent total volatile fatty acids (TVFA). This seems to indicate that, at higher OLR, the effluent total COD and mainly soluble COD is largely composed of the unused volatile acids produced in the reactor.

Given that the buffering capacity of the experimental system was found to be at favourable levels with excessive total alkalinity present at virtually all loadings, the efficiency of the process and the rate of methanogenesis was not very affected. The experimental data obtained in this work indicate that a total alkalinity of about 1.7 g/L as CaCO3 is sufficient to prevent the pH from dropping to below 7.0 at an OLR of 9.05 g COD/(L d) for the most concentrated substrate used (OMSW 2).

The pH in the reactor was always higher that 7.0 for all the HRTs and OLRs studied corresponding to the most diluted OMSW studied. In addition, pH values equal or higher than 6.9 were observed for OLRs lower than 12.02 g COD/(L d) and HRTs higher than 12.5 d when the most concentrated influent was processed, with pH of 7.2 as a maximum value achieved. This high stability can be attributed to carbonate/bicarbonate buffering. This is produced by the generation of CO2 in the digestion process which is not completely removed from the reactor as gas. Buffering in anaerobic digestion is normally due to bicarbonate, as carbonate is, generally, negligible if compared to the bicarbonate (carbonate/bicarbonate ratio is equal to 0.01 for pH 8.2) (Speece, 1983). The buffering guards

Influence of Substrate Concentration on the Anaerobic

from 12.02 to 15.03 g COD/(L d).

temperature is demonstrated.

0.00

influent. (■: OMSW 1; ●: OMSW 2).

**5.3 Kinetic evaluation** 

0.50

1.00

1.50

**r CH4 (L CH4/(L d))**

2.00

2.50

to the amount of substrate consumed, then:

Degradability of Two-Phase Olive Mill Solid Waste: A Kinetic Evaluation 87

This decrease in the methane production at the highest OLR values might be attributed to an inhibition of the methanogenic bacteria at high OLR values, which caused an increase in effluent TVFA contents and TVFA/Alkalinity ratio, as can be seen in Table 3. Specifically, TVFA content increased from 1.25 to 1.57 g/L (as acetic acid) when the OLR was increased

The experimental data listed in Tables 2 and 3 were used to determine the methane yield coefficient, *Yp*. As the volume of gas produced per day, *rCH4*, is assumed to be proportional

where *S0* and *S* are the substrate concentrations (expressed as g COD/L) at the digester inlet and effluent, respectively, and *q* is the feed flow-rate. By plotting Eq (1) in the form *rCH4* against *q (S0 – S)* (Figure 3), the following values of the methane yield coefficients with their 95% confidence limits were obtained for the two substrate concentrations used: 0.300 ( 0.001) and 0.200 ( 0.006) L methane STP/g COD removed when the OMSW 1 and OMSW 2, respectively, were processed. These values agree with the data reported in the literature for anaerobic treatment of food industry wastewaters (Borja et al., 1995; Maqueda et al., 1998; Martín et al., 1993). Taking into account that, theoretically, 0.35 L of methane is produced per gram of COD removed when the starting compound is glucose (Wheatley, 1990), the effectiveness of the anaerobic process in converting OMSW into methane at mesophilic

> 0.00 2.00 4.00 6.00 8.00 10.00 12.00 **q(So-S) (g COD removed/d)**

Fig. 3. Variation of the volume of methane produced per day, *rCH4*, as a function of the product of the differences of substrate concentrations at the reactor inlet (*S0* in g COD/L) and outlet (*S* in g COD/L) and the feed flow-rate (*q* in L/day) for the two OMSWs used as

Since the early 1980's, Stover and Kincannon have proposed a design concept of total organic loading rate and established a kinetic model for biofilm reactors. In this model the

*rCH4* = *Yp q (S0 – S)* (1)

against possible acidification of the reactor giving a pH of the same order as the optimal for methanogenic bacteria (Wheatley, 1990).

The TVFA/Alkalinity ratio can be used as a measure of process stability (Wheatley, 1990): when this ratio is less than 0.3-0.4 the process is considered to be operating favourably without acidification risk. As was observed in Tables 2 and 3 the ratio values were lower than the suggested limit value for OLRs lower than 9.05 g COD/(L d) in the experiments corresponding to the highest influent substrate concentrations studied (OMSW 2). For this substrate, between HRTs of 50.0 and 16.6 days, the TVFA/Alkalinity ratio was always lower than the above-mentioned failure limit and the TVFA values were always lower than 1,08 g/L (as acetic acid). However, at a HRT of 10.0 days, a considerable increase of the TVFA/Alkalinity ratio until a value of 0.95 was observed in the reactor, which was mainly due to a considerable increase in the TVFA concentration (1.57 g/L as acetic acid) with simultaneous decrease in alkalinity (1.32 g/L, as CaCO3).

#### **5.2 Influence of substrate concentration on the methane production rates and methane yield coefficients**

The volumetric methane production rates as a function of OLR are illustrated in Figure 2. As can be seen the volume of methane produced per day increased linearly with increased OLR up to OLR values of 3.45 and 12.02 g COD/(L d) for the influents OMSW 1 and OMSW 2, respectively. After these values a slight decrease was observed in the cases studied over the different ranges tested. Apparently, the activity of methanogenic bacteria was not impaired up to OLR values of 12.02 g COD/(L d) for the most concentrated influent (OMSW 2) used because of the appropriate stability and adequate buffering capacities provided in the experimental system. Nevertheless, the methane production rate decreased slightly from 2.12 to 2.05 L CH4/(L d) when the OLR was increased from 12.02 to 15.03 g COD/(L d).

Fig. 2. Variation of the methane production rate, *rCH4*, with the OLR (g COD/(L d)) of the reactor for the two OMSWs used as influents (■: OMSW 1; ●: OMSW 2).

against possible acidification of the reactor giving a pH of the same order as the optimal for

The TVFA/Alkalinity ratio can be used as a measure of process stability (Wheatley, 1990): when this ratio is less than 0.3-0.4 the process is considered to be operating favourably without acidification risk. As was observed in Tables 2 and 3 the ratio values were lower than the suggested limit value for OLRs lower than 9.05 g COD/(L d) in the experiments corresponding to the highest influent substrate concentrations studied (OMSW 2). For this substrate, between HRTs of 50.0 and 16.6 days, the TVFA/Alkalinity ratio was always lower than the above-mentioned failure limit and the TVFA values were always lower than 1,08 g/L (as acetic acid). However, at a HRT of 10.0 days, a considerable increase of the TVFA/Alkalinity ratio until a value of 0.95 was observed in the reactor, which was mainly due to a considerable increase in the TVFA concentration (1.57 g/L as acetic acid) with

**5.2 Influence of substrate concentration on the methane production rates and** 

The volumetric methane production rates as a function of OLR are illustrated in Figure 2. As can be seen the volume of methane produced per day increased linearly with increased OLR up to OLR values of 3.45 and 12.02 g COD/(L d) for the influents OMSW 1 and OMSW 2, respectively. After these values a slight decrease was observed in the cases studied over the different ranges tested. Apparently, the activity of methanogenic bacteria was not impaired up to OLR values of 12.02 g COD/(L d) for the most concentrated influent (OMSW 2) used because of the appropriate stability and adequate buffering capacities provided in the experimental system. Nevertheless, the methane production rate decreased slightly from 2.12 to 2.05 L CH4/(L d) when the OLR was increased from 12.02 to 15.03 g COD/(L d).

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 **OLR (g COD/(L d))**

Fig. 2. Variation of the methane production rate, *rCH4*, with the OLR (g COD/(L d)) of the

reactor for the two OMSWs used as influents (■: OMSW 1; ●: OMSW 2).

methanogenic bacteria (Wheatley, 1990).

**methane yield coefficients** 

0.00

0.50

1.00

**r CH4 (L CH4/(L d))**

1.50

2.00

2.50

simultaneous decrease in alkalinity (1.32 g/L, as CaCO3).

This decrease in the methane production at the highest OLR values might be attributed to an inhibition of the methanogenic bacteria at high OLR values, which caused an increase in effluent TVFA contents and TVFA/Alkalinity ratio, as can be seen in Table 3. Specifically, TVFA content increased from 1.25 to 1.57 g/L (as acetic acid) when the OLR was increased from 12.02 to 15.03 g COD/(L d).

The experimental data listed in Tables 2 and 3 were used to determine the methane yield coefficient, *Yp*. As the volume of gas produced per day, *rCH4*, is assumed to be proportional to the amount of substrate consumed, then:

$$r\_{\rm CH4} = Y\_p \, q \, (\text{S}\_0 - \text{S}) \tag{1}$$

where *S0* and *S* are the substrate concentrations (expressed as g COD/L) at the digester inlet and effluent, respectively, and *q* is the feed flow-rate. By plotting Eq (1) in the form *rCH4* against *q (S0 – S)* (Figure 3), the following values of the methane yield coefficients with their 95% confidence limits were obtained for the two substrate concentrations used: 0.300 ( 0.001) and 0.200 ( 0.006) L methane STP/g COD removed when the OMSW 1 and OMSW 2, respectively, were processed. These values agree with the data reported in the literature for anaerobic treatment of food industry wastewaters (Borja et al., 1995; Maqueda et al., 1998; Martín et al., 1993). Taking into account that, theoretically, 0.35 L of methane is produced per gram of COD removed when the starting compound is glucose (Wheatley, 1990), the effectiveness of the anaerobic process in converting OMSW into methane at mesophilic temperature is demonstrated.

Fig. 3. Variation of the volume of methane produced per day, *rCH4*, as a function of the product of the differences of substrate concentrations at the reactor inlet (*S0* in g COD/L) and outlet (*S* in g COD/L) and the feed flow-rate (*q* in L/day) for the two OMSWs used as influent. (■: OMSW 1; ●: OMSW 2).

#### **5.3 Kinetic evaluation**

Since the early 1980's, Stover and Kincannon have proposed a design concept of total organic loading rate and established a kinetic model for biofilm reactors. In this model the

Influence of Substrate Concentration on the Anaerobic

0.00

**6. Conclusions** 

**7. References** 

Spain, ISBN 84-7114-983-4.

52, No. 2, pp. 157-162, ISSN 0960-8524.

76, No. 1, pp. 45-52, ISSN 0960-8524.

0.20

0.40

0.60

0.80

**HRT/(So-S) (d L g-1 COD)**

1.00

1.20

1.40

Degradability of Two-Phase Olive Mill Solid Waste: A Kinetic Evaluation 89

0 10 20 30 40 50

Fig. 4. Determination of the kinetic parameters using the modified Stover-Kincannon model

The kinetic constants obtained define the bio-treatability of the two-phase olive mill solid waste. The values obtained for *Rmax* and *KB* were similar to those obtained for other substrates of high organic content. The increase in the maximum methane removal rate for the most concentrated two-phase olive mill solid waste used demonstrated the good adaptation of the bacterial inoculum used to the increase in the substrate concentration. This adaptation allowed the microorganisms to work with high stability even with high organic matter concentrations in the fed substrate. These results can be used to estimate the treatment efficiency of industrial-scale reactors working with similar operational conditions.

Alba, J.; Hidalgo, F.; Ruiz, M.A.; Martínez, F.; Moyano, M.J.; Borja, R.; Graciani, E. & Ruiz,

APHA (American Public Health Association). (1989). Standard Methods for the Examination of Water and Wastewater, 17th ed., APHA, Washington DC, USA, ISBN 087553161X. Borja, R.; Banks, C.J. & Wang, Z. (1995). Effect of organic loading rate on anaerobic treatment

Borja, R.; González, E.; Raposo, F.; Millán, F. & Martín, A. (2001). Performance evaluation of

M.V. (2001). Elaboración de aceite de oliva virgen. In: *El Cultivo del Olivo*, D. Barranco, R. Fernández-Escobar, L. Rallo (Ed.), 551-588, Mundi-Prensa, Madrid,

of slaughterhouse wastewater in a fluidised-bed reactor. *Bioresource Technology*, Vol.

a mesophilic anaerobic fluidized bed reactor treating wastewater derived from the production of proteins from extracted sunflower flour. *Bioresource Technology*, Vol.

for the two-phase OMSW 1 and OMSW 2. (■: OMSW 1; ●: OMSW 2).

**HRT (days)**

substrate utilization rate is expressed as a function of the organic loading rate by monomolecular kinetics for biofilm reactors such as rotating biological contactors and biological filters (Yu et al., 1998). This kinetic model can be used to describe carbonaceous removal in terms of BOD (biochemical oxygen demand), COD (chemical oxygen demand) and TOC (total organic carbon) as well as for nitrification.

The original Stover-Kincannon model (Kincannon and Stover, 1982) (Equation 2) was initially proposed for rotating biological contactor (RBC) systems and can be expressed by the following equation:

$$\mathrm{dS/dt} = \left[ \mathrm{R}\_{\mathrm{max}} \left( q \mathrm{S}\_{o} / A \right) \right] / \left[ \mathrm{K}\_{b} \mathrm{+} \left( q \mathrm{S}\_{o} / A \right) \right] \tag{2}$$

where: *A* is the disc surface area where the active biomass is attached; *S* is the substrate concentration in the reactor (in COD units) for a time (*t*); *So* is the initial substrate concentration; *q* is the flow rate; *Rmax* is the maximum removal rate constant and *KB* is the saturation value constant (in g COD/(L d)).

In the modified Stover-Kincannon model the substrate utilization rate is expressed as function of the organic loading rate as follows (Yu et al., 1998).

$$\mathbf{dS}/\mathbf{dt} = \left[\mathbf{R}\_{\text{max}}\left(q\mathbf{S}\_o/\mathcal{V}\right)\right]/\left[\mathbf{K}\_b\mathbf{+}\left(q\mathbf{S}\_o/\mathcal{V}\right)\right] \tag{3}$$

where *V* is the volume of the anaerobic reactor. The term d*S*/d*t* is defined for a steady-state relationship for different authors as:

$$\text{dS}/\text{dt} \equiv q \left( \text{S}\_o - \text{S} \right) / V \tag{4}$$

Linearization of equation (3) gives:

$$\mathbf{V} / \left[ \mathbf{q} \left( \mathbf{S}\_o \mathbf{-} \, \mathbf{S} \right) \right] = \left[ \mathbf{K}\_B \ \mathbf{V} / \left( \mathbf{R}\_{\text{max}} \, \mathbf{q} \, \mathbf{S}\_o \right) \right] + \left[ \mathbf{1} / \mathbf{R}\_{\text{max}} \right] \tag{5}$$

In continuously stirred tank reactors the hydraulic retention time (*HRT*) can be defined as: *HRT*=*V/q*, so equation (5) can be written as follows:

$$\text{(HRT)} / \text{(S}\_{o}\text{- }\text{-S)} = \left[\text{K}\_{\text{B}} \left(\text{HRT}\right) / \left(\text{R}\_{\text{max}}\text{ S}\_{o}\right)\right] + \left[\text{1}\left/\text{R}\_{\text{max}}\right] \tag{6}$$

According to this model a plot of (*HRT*)/(*So* - *S*) versus *HRT* should give a straight line of intercept [1/*Rmax*] and slope equal to *KB* /(*Rmax So*).

As can be seen in Figure 4 the experimental data fitted to a straight line with R2= 0.9992 for OMSW 1 and R2= 0.9999 for OMSW 2. The maximum removal rate constant (*Rmax*) increased from 26.6 to 83.3 g COD/(L d) when the OMSW concentration changed from 35 to 150 g COD/L, indicating a good adaptation of the initial inoculum to the OMSW treated and to increasing concentrations of organic matter fed. The saturation value constants (*KB*) were 27.7 g COD/(L d) and 82.7 g COD/(L d) for OMSW 1 and OMSW 2, respectively. The values of *Rmax* and *KB* obtained for the concentrated OMSW were similar to those obtained by other authors for the anaerobic digestion of soybean wastewaters (Yu et al., 1998) and molasses (Büyükkamaci & Filibeli, 2002). Stover and Campana (2003) have shown that in the model *Rmax* is reduced by refractory organics and toxicity. Moreover, the refractory compounds change *KB* significantly from *Rmax*. These affirmations are in agreement with the data obtained in these experiments, where the higher organic concentration of OMSW 2 gave *Rmax* values higher than for OMSW 1.

Fig. 4. Determination of the kinetic parameters using the modified Stover-Kincannon model for the two-phase OMSW 1 and OMSW 2. (■: OMSW 1; ●: OMSW 2).
